Very Long-Term Outcome After Stroke in Auckland, New Zealand
Background and Purpose— Limited information exists on the long-term outcome from stroke. We aimed to determine survival and health status at 21-year follow-up of patients who participated in a population-based stroke incidence study undertaken in Auckland, New Zealand.
Methods— During 12 months beginning March 1, 1981, half of all residents of Auckland with acute first-ever or recurrent stroke (n=680) were assessed and followed up prospectively during the next 2 decades. In 2002, their vital status and health-related quality of life (HRQoL) using the 36-item short-form questionnaire (SF-36) were determined by telephone interviews. Kaplan–Meier survival probabilities for the stroke cohort were compared with life table estimates for the New Zealand population. The SF-36 profile of 21-year stroke survivors was compared with a standardized New Zealand population.
Results— Overall, 626 of the original cohort had died and 4 were lost to follow-up, leaving 50 (7%) individuals (57% male; mean age 70 years) available in 2002, of whom 12% were residents of an institutional care facility and 19% required help with everyday activities. The stroke cohort had nearly twice the mortality rate of the New Zealand population, but the SF-36 profile of very long-term stroke survivors was broadly similar to the general population.
Conclusions— Because stroke is generally a disease of older people and has a high case fatality, it is not surprising that <1 in 10 people survive 2 decades after onset. However, of those who do, their HRQoL profile suggests that they meld relatively successfully within the general population, despite ongoing disability and a higher mortality risk.
The global burden of death and dependency associated with stroke is well established.1 However, much of our understanding about such outcomes from stroke is derived from studies with only a few years of follow-up2–4 so that data pertaining to the long-term prognosis after stroke are limited.5–14 Among such data sets is the Framingham Study, which has made significant contributions to knowledge of stroke outcome,13,14 including showing that survival rates and disability profiles of stroke patients are similar to those of controls at ≥20 years after first stroke.14 Yet, the unique nature of this cohort and the small numbers (only 10 people were included at the end of follow-up) limit the ability to generalize the data.
Because stroke is often part of a chronic disease process, with survivors facing high risks of recurrent vascular events and ongoing morbidity, it is important for health care planning, both preventative and therapeutic, that more information is obtained on long-term stroke outcome in different populations. In this study, we aimed to follow up all stroke patients who participated in the first Auckland Region Coronary Or Stroke (ARCOS) Study of 1981 to 198215 to (1) compare survival after stroke with life expectancy among individuals of the same age and sex in New Zealand; and (2) assess physical functioning and compare the health-related quality of life (HRQoL) of 21-year survivors with that of national age- and sex-standardized normal values.
Subjects and Methods
The ARCOS Study design has been described in detail.2,15 Briefly, a register was established to include 50% of all cases of acute stroke that occurred among all residents of the Auckland, New Zealand, population (829 545 according to the 1981 Census) during a 12-month period beginning March 1, 1981. Stroke was clinically defined according to the World Health Organization criteria of “rapidly developing signs of a focal (or global) disturbance of cerebral function leading to death or lasting longer than 24 hours, with no apparent cause other than vascular.”16 Multiple overlapping sources were used to ascertain cases, including prospective and retrospective searches of hospital and other medical records, death certificates, and autopsy reports, and by maintaining regular contact with all long-term residential care facilities such as rest homes and private hospitals. A random sample of 50% of all registered general practitioners in the Auckland area were asked to report all strokes that occurred among their patients to ensure that all individuals with acute stroke who were managed entirely outside of the acute hospital sector were included. All subjects with confirmed stroke had a baseline assessment undertaken by trained research nurses as soon as possible after notification, and the outcomes of the cohort were assessed at 1, 6, and 12 months, and at 3, 8, and 16 years after the index event. The study was approved by the Auckland Ethics Committee.
Baseline Assessment and Follow-Up
Baseline data collected through structured in-person interviews with subjects (or proxy) using standardized questionnaires included sociodemographic characteristics; risk factors, including smoking status and previous history of stroke, high blood pressure (hypertension), diabetes mellitus, and heart attack; and clinical features and management of the stroke. Cases with sufficient data were classified into the major stroke subtypes of cerebral infarction, subarachnoid hemorrhage (SAH) or intracerebral hemorrhage.
All subjects (n=74) not known to have died at the 16-year follow-up assessment in 1997, and who had agreed to subsequent contact, were included in the current study. Contact was initiated, principally by telephone, for subjects who were thought to be alive. After gaining informed consent, a structured telephone interview was conducted with either the subject or a reliable proxy, usually the spouse. For those people for whom previous contact details were out of date, exhaustive attempts were made to locate them using telephone directories, electoral rolls, and by contacting relatives, with final checks of vital status being made through the New Zealand Health Information Service for those people who could not be contacted.
At follow-up, information was collected on living arrangements and whether the subject required help with everyday activities of daily living (ADL). The 36-item short-form questionnaire (SF-36) was used to assess HRQoL. The SF-36 has been validated for use in stroke patients17,18 and comprises 36 questions covering 8 domains: physical functioning (PF), role limitations resulting from physical problems (RP), bodily pain (BP), general health (GH), vitality (Vt), social functioning (SF), role limitations attributable to emotional problems (RE), and mental health (MH). An absence of problems in a domain is indicated by a score of 100 for the PF, RP, BP, SF, and RE domains, and of 50 for the GH, Vt, and MH domains. For example, a score of 100 in PF indicates an ability to perform all activities without limitations because of ill health, whereas a score of 50 in MH indicates an ability to function without personal or emotional problems. To obtain scores of >50 for GH, Vt, and MH, a subject must rate his/her health positively, for example, a score of 100 in MH would indicate that the subject feels “peaceful and happy and is calm all the time.” Scores for the SF-36 are available for the New Zealand population, including Maori people, enabling age and sex comparisons with stroke survivors.
Differences in baseline characteristics between subjects who had died and those who survived for 21 years were examined with theχ2 test for categorical variables and a t test for continuous variables. The Kaplan–Meier product-limit method was used to generate survival P values and curves for the ARCOS cohort on the basis of date of death and censoring for individuals lost to follow-up using the date of last assessment. These data were compared with life table survival estimates for the general New Zealand population (>45 years old) on the basis of the 1981 Census and subsequent 5 yearly official mortality statistics, with adjustments made for the older age and different sex structure of the ARCOS Study population. SF-36 domain scores and SEs were calculated according to standard guidelines19,20 and analyzed using standard parametric methods. Crude mean SF-36 scores for stroke survivors were compared with age and sex standardized New Zealand normal values.20 All data were analyzed using SAS for Windows (SAS Institute)21 and reported with 95% CIs.
During the 21 years of follow-up, 626 (92%; 95% CI, 90% to 94%) of the 680 subjects registered in the ARCOS Study of 1981 to 1982 were known to have died. Only 4 (<1%) people were lost to follow-up, including 2 men and 2 people who had moved overseas and were not able to be contacted. This left 50 stroke survivors available for interview, although this was completed by a proxy (a close relative or caregiver) in 4 (8%).
Table 1 shows the baseline characteristics of all the ARCOS participants and of the 626 deceased subjects compared with the 54 survivors at 21 years. Very long-term survivors were significantly younger, had less premorbid dependency, were more likely to have smoked but less likely to have had diabetes, and their strokes were less disabling, more often a first-ever event, and attributable to SAH at onset.
Table 2 shows that of the 50 survivors interviewed (age mean±SD; 70±13 years; 56% male), more than half considered themselves to have made a full recovery from their stroke. Only one fifth of survivors, all of whom were previously independent, said they required help with ADL, which was considered by all but 1 to be caused by the effects of stroke. However, only 6 of the 9 dependent individuals were residents of a long-term institutional care facility; they tended to be older (mean age 78 years) and female.
Figure 1 shows the Kaplan–Meier estimates for the stroke cohort against the adjusted national life table estimates from 1981, with the curves continuing to diverge at the end of follow-up. On average, the stroke cohort had a 70% greater risk of dying than individuals of the same age and sex in the general population. The risk of death was greatest (double) during the first 5 years after the onset of stroke, attenuating to 17% above that of the natural mortality curve of the general population by 2001.
SF-36 data were available for 48 (96%) survivors and are summarized in Table 2. The crude mean scores for the GH, MH, and Vt domains were all >50, indicating that the group evaluated their health “positively.” Figure 2 shows the standardized mean SF-36 scores of survivors compared with the New Zealand population. With the exception of the Vt domain, there were no significant differences in SF-36 profiles. Because the standardized scores were based on noninstitutionalized people, we performed a second comparative analysis excluding the 6 institutionalized stroke survivors and produced similar findings (data not shown).
This population-based study confirms the high risk of death after stroke, with <1 in 10 people surviving for 21 years after illness onset, which was approximately double the mortality rate of the general population during the study period. Although the very high early case fatality and gradual decline in the risk of all-cause mortality thereafter is in agreement with previous population-based studies, our data also show that a history of stroke continues to predict a reduced life expectancy, even after 20 years. A more positive note was the finding of few people reporting dependency with a comparable SF-36 profile to New Zealanders of the same age and sex, suggesting an absence of major ongoing stroke-related disability in the small numbers of very long-term stroke survivors in Auckland.
A major strength of this study was the prospective population-based design that is regarded as having fulfilled most of the stringent criteria required for an “ideal” stroke incidence study22,23 and for an outcome study.24 In particular, the study used established clinical diagnostic criteria and standardized measures and procedures to ensure completeness in ascertainment of a representative sample of stroke cases in a large, well-defined population and the near-complete follow-up of cases from early to 21 years after stroke onset.
However, we acknowledge some limitations to the study. First, different methods were used for the survival analyses: Kaplan–Meier survival estimates for the ARCOS Study cohort and life tables for the New Zealand population because survival data were not available specifically for the population of Auckland. Despite our efforts to minimize any differences in the structure of these 2 populations by using life table estimates only for people ≥45 years old and by making further adjustments for the age and sex structure of the ARCOS Study population, there is the likelihood of bias being introduced in these analyses. However, we believe that any such bias is likely to be small because the sociodemographic structure of the Auckland and New Zealand populations as a whole were very similar in 1981, and both analyses included similar periods of inclusion and observation for individuals, thus taking into account the potential important period effects on survival.25 Although there are small numbers toward the tails of the curves, especially for the stroke cohort, the shapes of the curves are different and will not converge until all people have died in both cohorts, indicating that stroke patients continue to have a higher risk of death even after ≥20 years. Another limiting factor is the small number of subjects included in the assessment of HRQoL. Although there was more than 80% power to detect 10-point differences between groups across most of the 8 domains on the SF-36, the sample was too small to detect smaller but more clinically significant reductions in HRQoL among very long-term stroke survivors.
Our findings compare favorably with the limited number of other reports on the long-term prognosis of stroke. Cumulative case fatalities have been estimated at 40% to 60% at 5 years4,10,26 and >80% at 14 years after acute stroke in a hospital-based series from Linkoping, Sweden,8 and a community-based study in Helsinki, Finland.7 The latter study also found that nearly all long-term stroke survivors who continued to live at home were free of stroke-related disabilities. The study with the longest period of follow-up is from Framingham, Mass, where people who experienced strokes between 1972 and 1974 were assessed biennially for >3 decades.14 At the end of the follow-up period (mean time 27.4 years), only 9 (6% of 147) individuals remained alive; 8 were independent in ADL, and only 1 required institutional care. The higher proportion (19%) of survivors with ongoing dependency in ADL in our study may reflect differences in the assessment methods, but it may also be related to differences in case mix and the health care and social support systems in the 2 populations. Imprecision of the estimates because of the small numbers of subjects in both studies also limit the conclusions that can be drawn.
The population-based study that is most similar to the ARCOS Study is the Perth Community Stroke Study,12 which reported recently that approximately one fifth of stroke patients survived for 10 years, and 1-year survivors experienced an average annual case fatality of ≈5%, which was twice that of the population of Western Australia. We were not able to obtain cause-specific mortality data for the ARCOS cohort to confirm whether there is a similar high and ongoing risk of death from cardiovascular disease to that in the Perth study.
Survival is clearly the most fundamental measure of stroke outcome, the magnitude of the prevalent burden of stroke being influenced strongly in case fatality or life expectancy ratios to that of the general population. In the acute phase, the chance of surviving a stroke is generally a complex function of the extent of cerebral damage and the complications of disability, whereas in the longer term, age, sex, and continued exposure to vascular risk factors influence survival. To what extent positive trends in survival after stroke are related to advances in initial treatment and secondary prevention remains to be adequately elucidated.27
In conclusion, our data reaffirm the high early and ongoing risk of death among stroke patients, but they also suggest that functioning and HRQoL of those patients who survive ≥2 decades are comparable to that of the general population. These data are likely to reflect a better life expectancy for healthier people given certain favorable characteristics, including effects of illness at onset, among the very long-term stroke survivors of the ARCOS cohort. Thus, although it could be argued that stroke survivors are never truly free of effects of stroke, our data could also be viewed positively in the ability of many people to readjust, given sufficient time, to residual and often considerable stroke-related disability.
The first ARCOS Study, and subsequent follow-up studies, were supported by grants from the Health Research Council of New Zealand. The current study was made possible by the Auckland Medical Research Foundation, which funded a summer research studentship for Wallace Brownlee at the Faculty of Medicine and Health Sciences of the University of Auckland. We thank the ARCOS research staff for ensuring that the initial and follow-up assessments were complete and of high quality, and Jit Cheung and Martin Tobias of the Ministry of Health, New Zealand, for help with statistical analyses.
- Received February 22, 2004.
- Revision received April 4, 2004.
- Accepted May 2, 2004.
Murray CJL, Lopez AD. Global Health Statistics. Geneva, Switzerland: World Health Organization; 1996.
Bonita R, Ford M, Stewart A. Predicting survival after stroke: a three-year follow-up. Stroke. 1988; 19: 669–673.
Anderson CS, Jamrozik KD, Broadhurst R, Stewart-Wynne EG. Predicting survival among different subtypes of stroke: experience from the Perth Community Stroke Study. Stroke. 1994; 25: 1935–1944.
Brønnum-Hansen H, Davidsen M, Thorvaldsen P; Danish MONICA Study Group. Long-term survival and causes of death after stroke. Stroke. 2001; 32: 2131–2136.
Scmidt E, Smirnov V, Ryabova V. Results of the seven-year prospective study of stroke patients. Stroke. 1988; 19: 942–949.
Indredavik B, Bakke F, Slordahl SA, Rokseth R, Haheim LL. Stroke unit treatment: 10-year follow-up. Stroke. 1999; 30: 1524–1527.
Tuomilehto J, Nuottimaki T, Salmi K, Aho K, Kotila M, Sarti C, Rastenyte D. Psychosocial and health status in stroke survivors after 14 years. Stroke. 1995; 26: 971–975.
Engstad T, Viitanen M, Arnesen E. Predictors of death among long-term stroke survivors. Stroke. 2003; 34: 2876–2880.
Dennis M, Burn J, Sandercock P, Bamford J, Wade D, Warlow C. Long-term survival after first-ever stroke: the Oxfordshire Community Stroke Project. Stroke. 1993; 24: 796–800.
Hackett ML, Duncan JR, Anderson CS, Broad JB, Bonita R. Health-related quality of life among long-term survivors of stroke: results from the Auckland Stroke Study, 1991–1992. Stroke. 2000; 31: 440–447.
Hardie K, Hankey GJ, Jamrozik K, Broadhurst RJ, Anderson C. Ten-year survival after first-ever stroke in the Perth Community Stroke Study. Stroke. 2003; 34: 1842–1846.
Sacco R, Wolf P, Kannel W, McNamara P. Survival and recurrence following stroke. The Framingham Study. Stroke. 1982; 13: 290–295.
Gresham GE, Kelly-Hayes M, Wolf PA, Beiser AS, Kase CS, D’Agostino RB. Survival and functional status 20 or more years after first stroke: the Framingham Study. Stroke. 1998; 29: 793–797.
Bonita R, Beaglehole R, North JD. Event, incidence and case fatality rates of cerebrovascular disease in Auckland, New Zealand. Am J Epidemiol. 1984; 120: 236–243.
Anderson C, Laubscher S, Burns R. Validation of the short form 36 (SF-36) health survey questionnaire among stroke patients. Stroke. 1996; 27: 1812–1816.
Dorman P, Slattery J, Farrell B, Dennis M, Sandercock P. Qualitative comparison of the reliability of health status assessments with the EuroQol and SF-36 questionnaires after stroke. Stroke. 1998; 29: 63–68.
Ware JE Jr, Snow KK, Kosinski M, Gandek B. SF-36 Health Survey: Manual and Interpretation Guide. Boston, Mass: New England Medical Center, Health Institute; 1993.
Ministry of Health. Taking the Pulse: The 1996/97 New Zealand Health Survey. Wellington, New Zealand: Ministry of Health; 1999.
The SAS System for Windows Version 8. Cary, NC: SAS Institute Inc; 1999.
Sudlow CLM, Warlow CP. Comparing stroke incidence worldwide: what makes studies comparable? Stroke. 1996; 27: 550–558.
Sackett DL, Haynes RB, Guyatt GH, Tugwell MD. Clinical Epidemiology: A Basic Science for Clinical Medicine. 2nd ed. Boston, Mass: Little, Brown and Company; 1991: 173–185.
Hannerz H, Lindhardt M. Life expectancies among survivors of acute cerebrovascular disease. Stroke. 2001; 32: 1739–1744.
Hankey GJ, Jamrozik K, Broadhurst RJ, Forbes S, Anderson CS. Five-year survival after first-ever stroke and related prognostic factors in the Perth Community Stroke Study. Stroke. 2000; 31: 2080–2086.
Terent A. Trends in stroke incidence and 10-year survival in Söderhamn, Sweden, 1975–2001. Stroke. 2003; 34: 1353–1358.